Professionalized political avatars? A study of the twitter profiles of Dutch, Norwegian and Austrian MPs

Professionalized Political

Avatars?

A Study of the Twitter Profiles of Dutch,

Norwegian and Austrian MPs

University: Radboud University Nijmegen

Student: Eirin Vikki Kofoed, s4172221, eirinkofoed@hotmail.com

Master program: Comparative Political Science, Nijmegen School of

Management

Date submitted: 12 August 2015

Abstract

Social media has become increasingly popular the last years, with ever more people taking these online services in use. Also politicians are signing up on Social Networking Sites (SNS) in order to reach out to more voters. At the same time, it has been claimed that political communication has become more professionalized and personalized the last decades. This study focuses on

professionalization in the Political Avatars, or social media profiles, of Dutch, Norwegian and Austrian MPs. Although a growing amount of research is investigating the social media use of politicians, most have focused on checking which politicians are online and how active they are. This quantitative study, on the other hand, investigates the level of professionalization and what influences this in the Twitter profiles of the MPs of the most similar system cases of the Netherlands, Norway and Austria. The findings show that older MPs have less professionalized Twitter profiles than younger MPs, that MPs from populist parties have less professionalized profiles, and that Dutch MPs have more professionalized Twitter profiles than Norwegian and Austrian MPs. Several explanations for the different levels of professionalization in the Twitter profiles were presented, as well as recommendations for future research.

TABLE OF CONTENTS

INTRODUCTION

5

CHAPTER 1: THEORETICAL FRAMEWORK: PROFESSIONALIZATION

AND PERSONALIZATION 9

1.1. Societal developments: weakening social cleavages, low trust in parties and low engagement in politics……….…………9

1.2. The personalization of politics………...…..10

1.3. Professionalization of political communication, mediatization and personalization………..12

1.4. Conceptual unclarities: Professionalization and personalization…….………15

1.5. Characteristics of social media………....16

1.6. Dependent variable: professionalization in social media profiles………...18

1.7. Independent variables: individual and contextual variables: Hypotheses...………....18

1.7.1. Individual variables………..18

1.7.2. Contextual variables: meso- and macro-variables………...20

CHAPTER 2: METHODOGICAL FRAMEWORK 24

2.1.

2.1.1. Most similar systems design and consensus democracies……….……..24

2.1.2. The role of electoral systems for personalization………25

2.1.3. The role of the media system for professionalization………..………27

2.1.4. Characteristics of the three case countries – parties in parliament and social media penetration………....29

2.2. Operationalization………31

2.2.1. Measurement of the independent variables………..31

2.2.2. Measurement of the dependent variable professionalization: an advanced political avatar and a personal political avatar……….…..31

2.2.3. Measurement of the dependent variable – an advanced political avatar………..32

2.2.4. Measurement of the dependent variable – a personal political avatar……….33

2.3. Explanatory analysis………...…….34

2.3.1. Indices……….….35

2.3.2. Regression analyses……….39

CHAPTER 3: RESULTS: PROFESSIONALIZED POLITICAL AVATARS? 40

3.1. Research method………..………40

3.2. Descriptive analyses………..…..40

3.3. Bivariate and multiple OLS regression………..…………..41

3.3.1. Bivariate regressions………...….41

4.2.2. Multiple regression – Results and hypotheses……….44

3.3. Discussion………49

CHAPTER 4: CONCLUSION 54

4.1. Summary………..………54

4.2. Recommendations for further research………57

APPENDICES 65

Tables

Table 2: Operationalization of the independent variables………31

Table 3: Variables in the advanced user index………..…37

Table 4: Variables in the personalized user index……….37

Table 5: Variables included in the professionalized index………38

Table 6: Bivariate regressions for the three indices measuring professionalization……….….42

Introduction

Political communication and political campaigns are domains that have undergone great changes the last decades. With technological developments, the reduced role of social cleavages, weaker parties, and greater voter volatility, political communication and the media picture have changed dramatically the last six decades (Vliegenthart 2012). These developments are said to be part of the

professionalization of electoral campaigning (Strömbäck 2007). Professional campaigning is when marketing concepts are used when conducting electoral campaigns (ibid., 55) Among these

developments are an increased focus on the individual politician and especially the party leaders (increased personalization), increased negativity in politics and commercialization of politics, increased use of telemarketing, opinion polls, and particularly the use of specialized communication experts (Vliegenhart 2012, Strömbäck 2007). Among these new developments are also the use of Internet campaigning in the form of websites and social media, and the use of professional experts in managing online political communication.

At the same time, Social Network Sites (SNS) have become increasingly popular the latest years, with a large part of the population of Western democracies using these virtual services (Statista, the Statistics Portal 2015). Especially young people are frequent users of social media. For instance, in one of the cases used for in this study, the Netherlands, 93 % of people between the age of 12 to 18, and 98 % of people aged 18 to 25 use social media (CBS Statistics Nederland 2013). Politicians have started to use these sites, obviously seeing the potential they represent in promoting themselves, gaining popularity and ultimately attaining more votes at elections (Kruikemeier 2014, Svensson 2013). Social media has in this way been introduced as a strategy in electoral campaigning and in the promotion of political candidates. By using social media, they can (presumably) reach out to more people, and to other kinds of people than if they would only make use of traditional mass media (Spierings and Jacobs 2014, Svensson 2013, 209). However, despite the growing popularity of social media among politicians, little (but growing amounts of) research has been done on this subject compared to research on traditional political communication (Spierings and Jacobs 2014, 2).

Whereas many studies on political social media use up until now have focused on counting how many and which politicians are present on SNS (Vergeer et. Al. 2011, Grant et. Al. 2010, Hsu & Park, 2011), how many followers they have (Vergeer et. Al. 2011, Hsu & Park 2011), and how many

tweets or posts they send out (Vergeer et. Al., 2011, Grant et. Al. 2010, Hsu & Park 2011), and some research has been done on the content of their posts/tweets (Golbeck et. Al. 2010, Jackson and Lilleker 2011), how politicians present themselves on their profiles, on the other hand, has not been the focus of much research. The latter point is the aim of this master thesis. My assumption is that we can find characteristics of professionalized campaigning also in the social media use of politicians, following the trends of traditional media and politics. For example, putting out personal information and pictures about themselves will (possibly) enable politicians to “stand out from the crowd” and seem different from other politicians, and enable them to form emotional bonds with the voters (Vergeer et. Al. 2011). This master study aims at examining whether the social media profiles of politicians, or political avatars, display features of professionalized campaigning, such as personalization, but also whether the profiles is professionalized as in non-amateurish, i.e. the profile has consistent and “professional” design. This thesis also aims at determining what might explain the different levels of professionalization of politicians’ social media profiles. Is this influenced by the party the MPs belong to, or is this influenced more by their individual characteristics such as age, gender or ethnicity? The overall research question of this thesis is thus: What influences the level of professionalization of politicians’ political avatars?

This is a quantitative study which will look at how Dutch, Norwegian and Austrian politicians present themselves on Twitter by looking at their profile picture, background picture and profile information, trying to find similarities and differences between the MPs in these three countries. The research in this study is conducted by performing a Political Avatar Analysis (PAA) of the Twitter profiles, or political avatars, of Dutch, Norwegian and Austrian MPs. A Political Avatar Analysis is a frame for analyzing how politicians behave online and what kind of digital persona they want to portray themselves as (Radboud University Nijmegen 2015). This is done by quantitatively and qualitatively analyzing tweets, performing interviews and looking at politician websites and social media profiles (ibid.). A PAA is thus suitable for finding the characteristics of the usage of social media by politicians due to its holistic look at many aspects of this political communication (ibid.). Regression analyses will be performed in order to determine which factors influence the level of professionalization.

I have chosen to study the cases of Norwegian, Dutch and Austrian MPs’ Twitter profiles due to the similarities in these countries’ political and media systems. In this way, they are most-similar-system cases (Seawright and Gerring 2008). I take Josep Colomer’s (2011) framework on electoral systems and personalization as my departure point and work from the premise that what kind of electoral system a country has will determine the level of personalization in the political system. I assume the level of personalization of a political system will reflect itself in the social media use of politicians. If the level of personalization in these countries is expected to be low, I expect it to also be low in their Twitter profiles, if it is expected to be high, I expect it to be high in their Twitter profiles. I

also take Hallin and Mancini’s (2004) media system typology as a departure point, as the kind of media system a country has might influence social media use. The Netherlands, Norway and Austria are all examples of democratic corporatist media systems, another factor that can influence the level of professionalization of the politics in these countries. As the countries have similar electoral and media systems, I expect them to display similar levels of professionalization in their political and media system, and thus also in the politicians’ political avatars. In the following section, I will outline the relevant theories on political social media use, and on personalization and professionalization of politics and political communication.

Chapter 1: Theoretical Framework: Professionalization and

personalization

1.1 Societal developments: weakening social cleavages, low trust in parties and low engagement in politics

Western democracies have experienced major political changes the last decades. Robert Putnam (1995) pointed to the decline in civil engagement and social capital in the US, arguing that it has been devastating for civil political participation as well as for interest and trust in politics, a development possibly paralleled in many European countries. People have become less engaged in party politics, reflected in low voter turnout (Gray and Caul 2006, 1092), decrease in party membership (Mair and Van Biezen 2001) and political mobilization by parties (Dalton 1984). Another change can be seen in the very makeup of society, also profoundly influencing politics. Up until a couple of decades after the Second World War, Western societies were characterized by social cleavages that could easily be distinguished from each other economically or culturally and who had their own interests and ideological preferences and their own political parties (Vliegenthart 2012, 137). Voter volatility was low as all members of a cleavage usually voted for “their” party. If you were a manual worker, you voted for the Socialist or the Labor Party, if you were a business or factory owner, you voted for the Conservative or the Liberal Party, etc. This was the “heyday” of political parties as they had a solid and stable voter base and party membership, they were agenda setters, and journalists usually followed their lead (ibid.).

However, from the 1960’s, societal developments such as secularization and modernization led to social cleavages disappearing or becoming less important. This changed politics and how politicians communicated with the electorate (ibid.). People no longer automatically voted for “their” party, and voter volatility increased and partisanship decreased (partisan dealignment). Ideology was proclaimed dead (Bell 1960, as paraphrased in Karvonen 2009, 1), and parties changed from mass class parties to catch-all parties (Karvonen 2009, 1). Other developments such as political elites distancing themselves from normal citizens has also led to a decline in trust in politicians and parties and to a crisis of legitimacy of political parties and of institutions (Spierings and Jacobs, 2012, 3). Dalton and Weldon (2005) even argue that the decreasing trust in political parties pose a threat to democracy, as political parties are seen as necessary for representative democracy. When trust in

political parties is low, this also affects how people feel democracy works in their countries (ibid.). This might lead to increased support for extreme populist parties, to a fragmentation of the party system, to more cases of divided government, to increased elite recruitment, to less grass-root mobilization, etc (Dalton and Wattenberg, 2000).

1.2. The personalization of politics

At the same time, many argue that we have seen a process of personalization of politics in Western democracies (McAllister 2007, Dalton and Wattenberg 2000), perhaps as a response to the

abovementioned changes. Personalization of politics implies that individual politicians have become more important at the expense of political parties (Karvonen, 2009, 4). Karvonen (ibid., 2) argues that with weakening social structures and collective identities, and with loyalties becoming less important, the individual politician has become more important in determining “how people view politics and how they express their political preferences”. People vote increasingly for persons, not for parties (Manin, 1997, 219, as paraphrased in Karvonen, 2009, 2). Colomer (2011) similarly argues that personal accountability has become more important at the expense of partisan accountability, and represents a “neglected dimension” of politics. A personal “bond” between the individual politician and the voter is what connects them, instead of partisanship and group loyalties like before.

Karvonen (2009, 5) argues that personalization entails the following changes: firstly, institutions may stress individual politicians over collectivities more than in the past, secondly, the way politics is presented to the citizens through media and in electoral campaigns tends to stress the role of individuals politicians, thirdly, people come to see politics as a competition between individual politicians rather than organized collective interests, fourthly, people increasingly form their political preferences on the basis of their images of individual politicians, fifthly, people increasingly vote based on preferences formed on the basis of their evaluations of individual politicians, sixthly, the choices citizens make on the basis of their evaluations of individual candidates can decide which politicians or parties win elections, and seventh, power relationships in politics and society may come to be decided on the basis of the individual characteristics of politicians. This development might have come about as a result or consequence of the abovementioned developments, or it may have been suggested as a conscious strategy by political elites, scholars and pundits (Spierings and Jacobs 2012). However, it must be noted that the personalization thesis has also been questioned, as some scholars have found weak evidence, or no evidence at all, of such a development (Karvonen 2009, Aarts et. al. 2011, Kriesi 2010).

Hermansson (2011) outlines three general theoretical perspectives which interpret

personalization according to how voters regard themselves. The first, unaffiliated voter perspective, assumes that people feel little or no group belonging compared to what they used to feel, and that they feel politics does not concern them or is uninteresting, and do not identify with any party (ibid., 3).

Traditional identities are weakened, and the bond between politicians and voters erodes, but

personalization of politics might lead to voters feeling there is at least something that makes politics interesting. Voters will increasingly focus on packaging and presentation of political messages. It is likely that party leaders will become more important with voters with weak party identification (ibid., 4). This corresponds well to the developments outlined previously.

The second perspective sees voters as clients in a marketplace (ibid.3, 5). Here, parties are viewed as service-providers, and if these parties become too much alike, or cannot provide the voters with viable, clearly distinguishable alternatives, the voter will make a choice according to “external packaging” (ibid., 3). Examples of such external packaging could be superficial factors like the personalities or looks of the politicians instead of the content of their politics. The median voter theory assumes that most parties will tend to converge towards the center of the left-right continuum to appeal to the median voters, as most voters are situated in the middle of the specter (ibid., 5). In this way, the parties will become very much alike, providing incentives for voters to focus more on party leaders.

The third perspective is the presidentialization of politics. Some have argued that we see a presidentialization of politics. The latter is defined by Poguntke and Webb (2005, 1) as “a process by which regimes are becoming more presidential in their actual practice without, in most cases, changing their formal structure, that is, their regime-type”. A key feature of presidential systems is the

separation of the legislature and the executive, meaning that the executive is independent of, and not responsible to, the legislature (ibid., 2-3). Another feature is that the president is elected directly by, and responsible only to the electorate. The last key feature is that as the president is elected directly, and he has authority to hire and fire the members of the cabinet and they are directly responsible to him. In a parliamentary system, on the other hand, the executive is formed by the legislature and is responsible to the legislature, and there is a collective executive responsibility, meaning the executive as a whole emerges from the legislatures, including prime minister and cabinet ministers (ibid., 3).

Poguntke and Webb argue that a move towards presidentialization includes changes in three areas: the executive face, the party face and the electoral face (ibid.). When it comes to increasing power and autonomy in the executive face, we see that the leader achieves greater power and

autonomy vis-à-vis their parties. The second arena of change is the party face, and here we see a shift in internal power to the leader of the party. The third arena, the electoral face, we see an increased control of the leader and “a growing emphasis on leadership appeals in election campaigning” and that “the media coverage of politics focuses more on leaders”. Thus, we see an increased personalization (ibid., 10). The presidentialization thesis has, like the personalization thesis, been criticized as some researchers have not found strong evidence for this development (Karvonen 2010, Bakvis & Wolinetz 2005, Paloheimo 2005).

1.3. Professionalization of political communication, mediatization and personalization Karvonen (2010, 6) suggests there are three areas of research that focus on personalization, partly overlapping with the abovementioned theories. The first is studies of presidentialization, the second is studies of party leader effects in elections, and the third are studies with varying empirical foci. I would argue that research that focus on personalization in media is an important field of study. Media has been very important in conveying political messages and bringing politics out to the people ever since the beginning of politics, and has certainly not become less important the last decades.

Technological change has allowed media to penetrate bigger consumer markets in different ways than before, which has also profoundly shaped politics. It has been suggested that media has had a big part to play in making politics more personalized (McAllister 2007, 6, Karvonen 2010, 21), or has even been singled out as the most important cause for personalization (Swanson and Mancini 1996, as paraphrased in Karvonen, 2009, 4).

The way politicians communicate with the electorate has undergone great changes the last decades. There is talk of “postmodern” or “professional” campaign styles or strategies as mentioned by Vliegenthart. Blumler and Kavanagh (1999) outline several developments that have led to the changes in political communication the last decades. Modernization is one of these, which entails an increased “social differentiation and specialization, fragmenting social organizations, interests, and identities proliferating diverse lifestyles and moral stances; and fueling identity politics” (ibid., 210). A second is secularization, already mentioned above, which reduces the importance of denominational parties and party identification (ibid.). A third development mentioned by the authors is

economization, where economic factors and values become more important in politics (ibid.). As media markets become more saturated, the competitiveness in the market increases, and politicians and parties must increasingly fight for a place in the spotlight. Politics become more like

“infotainment” in order to catch the interest of the public. A further development is aestheticization, i.e. people become more preoccupied with stylishness, image and presentation, leading to politics becoming more associated with popular culture (ibid.). In addition to this we also see an increased rationalization (of administration and organizations), and mediatization.

Mediation of politics according to Strömbäck (2008, 230) is “a situation in which the media have become the most important source of information and vehicle of communication between the governors and the governed”, i.e. when the public depends primarily on the media to be informed about politics. This is not a new phenomenon, but people are relying more on the media than before for these purposes (ibid.). The term of mediatization refers to “the adaptation of politics to the media’s

rules” (Campus 2010, 220), where politicians increasingly adapt their behavior according to “what the media wants” when it comes to timing, location, framing of messages etc.

Vliegenthart (2012) in his study provides an overview of changes in political communication the last decades, a series of developments that he calls the professionalization of political

communication, which is the term I will use in this study. The first period, until around the 1960’s, he calls the period of “political logic”. In this period, the political parties set the agenda, they were dominant actors in political communication as journalists usually followed their lead, and their campaign efforts were usually directed towards getting the people of their own social cleavage out there to vote for them (ibid., 137). Political communication was essentially ideologically driven, with more focus on the party than on the party leader and individual politicians. With the decline of social cleavages and the rise of television, electoral campaigns became more important as politicians could gain or lose more in the period before the election. The rise of television made individual qualities such as good looks and strong debating skills more important. Political conflict also became more important at the expense of real political issue coverage (ibid., 138). Vliegenthart (ibid.) follows Brants and van Praag (2006) and calls this period the period of “public logic”, pointing to the increased importance of journalists and their concern for the public good, acting as “democratic watchdogs”. There was less focus on ideology, more focus on the party leader, and an increased use of professional political consultants in this period. The third period, the postmodern period, or the period of “media logic”, started in the 1990’s. Many of the changes that happened in the previous period were intensified or were extended in this period. We can, according to Vliegenthart (ibid.), see “increasing levels of electoral volatility, personalization, and negativity, whereas public opinion polls … become a very important guide of journalists, politicians, and policy makers”. Media has now taken the lead that was previously held by political parties (therefore the term “media logic”), and we see increasing levels of commercialization and mixing of entertainment with politics.

The term professionalized campaigning, according to Strömbäck (2007, 54), is when campaigns are

[…] characterized by being permanent, although with varying intensity; by the central campaign headquarters being able to coordinate the messages and the management of the campaign; and by using expertise in analyzing and reaching out to members, target groups and stakeholders, in analyzing its own and the competitors’ weaknesses and strengths and making use of that knowledge, and in news management.

Thus, most of the activities in professionalized electoral campaigning require the use of professional and specialized experts in managing campaigns. Gibson and Römmele (2001, 31-32) point to four important changes in electoral campaigning we can observe in the professionalized campaigns: firstly,

political parties have adopted new tools for campaigning such as electronic communication through Internet, direct mailing and telephone banks. Secondly, we see an intensification of already existing methods such as the use of opinion polls and focus groups. Thirdly, there has been a shift in power within parties, with the party leaders attaining more power, and externally by putting greater responsibility in the hands of media consultants. Fourthly, electoral campaigning has become more business-like as voters are seen as consumers and parties as service-providers providing the consumers with products (policies) to choose between through advertisement-like strategies.

It is unlikely, however, that all countries and all parties exhibit the same degrees of

professionalization of their campaigns. Gibson and Römmele (ibid., 34) point to external, systemic-level explanations such as the frequency of elections, the rules for political advertising and donation, and levels of party attachment in the electorate in explaining the move to professionalized political communication. They also point to more party-centered explanations (ibid., 36). For example, they point to major political change in parties as a response to external shocks (like low electoral support, new government coalitions) as an important explanatory factor. Certain organizational features such as party resources, internal structure and ideology are thought to make a transition to professionalized campaigning more likely (ibid., 37). In this argument, parties with greater resources, parties with a top-down structure, catch-all parties and right-wing parties are thought to be more likely to adopt to professionalized campaign strategies.

Strömbäck (2008, 238) outlines a situation where media has become more important, and where the media logic includes focusing on conflict and personalization. McAllister (2007, 7) argues that media focuses on personalities because it is easier to convey political messages and information through persons rather than through something abstract like a party or a party document. Particularly by using political leaders is television able to catch and hold the attention of the viewers. A further way that media projects political personalities is through political priming, i.e. “the process by which leaders are evaluated by voters, based on a leader’s performance on the issues that are considered to be of importance to voters” (ibid., 8). TV plays an important role in determining how this priming happens by influencing how the issue is framed in media, and by influencing which issues become politicized.

McAllister (ibid., 7) points to the fact that also political parties are to blame for these

developments, as they find it easier to market political choices to voters through personalities. Parties thus use media to their own advantage, they are not passive “victims” of the media. Campus (2010) points to the fact that politicians use media to create favorable images of themselves. Media images are powerful symbols that identify a politician and allow politicians to be present to people even if they are not physically present (ibid., 221). One strategy that can be used in creating an image of a politician is to “go personal” in that details about the private life of the politicians are revealed to the

public,for example through nonpolitical talk shows, thereby making the image of the politician more “humanized” (ibid.).

Developments in societal structures and in media (and perhaps partly of their own doing) have thus pushed politicians to actively seek out ways to reach out to the public more easily and in a personal way, and this is where social media comes in. Social media provides just such a tool that allows politicians to build their image, reach out to voters, and “be personal”.

1.4. Conceptual unclarities: Professionalization and personalization

Are professionalization and personalization two separate concepts, or the same? Or is one the

consequence of the other? Do they belong together at all, or should they be used separately? It is easy to see how one could become confused about the concepts of professionalization and personalization, as researchers do not seem to be consistent about their content and uses. This is what this section will try to clarify.

When it comes to the concept of professionalization, Negrine and Lilleker (2002) argue that the concept is used very vaguely, pointing to the fact that different concepts such as

“professionalization of politics”, “political marketing”, “campaign professionalism”, “source professionalization”, “Americanization” of political campaigns, etc, are all used to signify how political communication today is different from how it used to be. The term professionalization also differs depending on which field of study you look at. Firstly, political scientists often see

professionalization as “the ways in which politics has now become a full-time paid career” (Rush 1989, 2001, as quoted in Negrine and Lilleker 2002, 310). In political communication,

professionalization can mean hired media experts by parties, but can also refer to those who can manage the media, which can also be politicians themselves. It thus refers to a “professional” way of handling media, a set of skills to use modern media techniques (Schlesinger and Thumber 1994, 84, as paraphrased in Negrine and Lilleker 2002, 310). Professional here means the opposite of amateurism. Negrine and Lilleker (ibid., 307) suggest that professionalization may only refer to the way political actors have adapted to new technical innovations and to the specialization of tasks that we can see in all modern societies. In this process, parties have thus hired media experts, and the politicians themselves have become more media-savvy.

And how does personalization fit in here? Until now I have spoken of them as if they are two different developments while in fact, they are related. As mentioned earlier, Vliegenthart (2012) mentioned personalization as a part of the developments of the period of “media logic”, which again is a part of the process of professionalization. Holtz-Bacha (2002, 24), following Schultz (1998), mentions that professionalization of political communication (or “modernization” as he calls it) entails

a de-ideologization, i.e. ideology becomes less important while a focus on the individual politician, a process of personalization, becomes more important. Also Tenscher et. al. (2012, 148) mention personalization as a development of professionalization, or characteristic of professionalization, arguing that the more personalization (among other factors) is included in the campaign, the more professionalized the campaign is. Thus, personalization can be considered a part of

professionalization. This study will focus on the level of professionalization in social media, particularly on the level of professionalization in Twitter profiles of MPs.

1.5. Characteristics of social media

As already mentioned, the use of social media among the population in Western countries has increased, presumably also making politicians adopt social media profiles in order to reach more voters. Social media can be defined as “online communication platforms where the social seems to refer to the possibility of users to influence and interact with the content and each other in some way or another” (Svensson 2013, 2). Examples of social media can be SNS (Social Networking Sites) like Facebook, webblogs, microblogging sites such as Twitter, and wikis (Stieglitz and Dang-Xuan 2012, 1279). Social media is a unique possibility for politicians to reach out to citizens as it is cheap, anyone can make a social media profile, and it does not have gatekeepers like journalists as there are in traditional media (Spierings and Celis 2014). Politicians thus have a unique possibility of reaching out to a (potentially) immense audience at a low cost, including groups they would otherwise not reach that easily, such as people with low political capital (ibid.). In addition to this, social media allows for non-mainstream profiling and for orientation of in-groups (ibid.). In this way, politicians, single persons or organizations can mobilize people to rally for a certain cause. Social media thus has the opportunity of making more people politically involved (Vergeer et. al. 2011, 479).

A prominent characteristic of social media is that it is social, meaning there is social contact between two or more people, making it personal almost by definition, as it consists of the profiles of individual persons (although also organization pages for, for example, political parties). This media fits in well with the abovementioned developments, such as the increased focus on individual politicians (personalization) (Enli and Skogerbo 2013, 759). Social media allows politicians to show more personal sides of themselves, such as their family life and their personal likes and dislikes, but allows them to show their public lives as well (ibid.). Kruikemeier (2014) also points to

personalization as an important characteristic of online communication. Online communication will be focused on the individual politician, thus is personalized, and less focused on the party and

(privatization), emotions and feelings (emotionalization) or competencies they want to highlight (individualization) (ibid.). By being personal, parties get a face and voice, making it easier to communicate with the voters (ibid.).

Another important characteristic of social media is interactivity, i.e. you engage in reciprocal communication with someone, thereby directly reaching out to voters and perhaps forming emotional bonds with them. Being interactive with citizens can thus also be a way of being personal. On social media, citizens can receive answers to their questions posed on politicians’ profiles, and react to politicians’ tweets. Social media has been hailed as a way to promote participation and deliberation among citizens in politics and of democracy (ibid.). Due to their interactivity, it allows citizens to contact politicians and to have a dialogue with them. It does not cost citizens or politicians much time or resources to be involved in social media, making it a great potential for more participatory and interactive political institutions (ibid.). Kruikemeier (2014) in her article assumes that interactivity is thought to have a positive influence on people’s votes. This suggests that politicians’ avatar usage should be more interactive in order to attract votes. Kruikemeier et. Al. (2013) find that the use of personalization and interactivity in online political communication by politicians have positive effects on the political engagement of citizens. However, there are studies that show that politicians are not using this potential as much as could be expected, as for example Svensson’s (2013) study showed.

Svensson (2013, 206) outlines the reasons politicians might have to use Social Networking Sites in election campaigning. He outlines three reasons as the most important: firstly, to target voter groups (an instrumental rational approach), secondly, for deliberations with their constituency (a communicative rational approach), and thirdly, to make an image as politicians (an expressive rational approach to social media). Enli and Skogerbo (2013, 763-4) outline the following three motives for using social media: firstly, a marketing purpose to increase visibility of candidates and parties, i.e. it is used as another way to promote themselves. The second motivation for using social media is the mobilization of the people, for example in events, demonstrations and to vote. A third motivation is the desire to have a dialogue with the voters. Svensson’s first and second reasons overlap to a great extent with Enli and Skogerbo’s first and third motivation, and we can also observe Campus’ argument that politicians use the media to create favorable images of themselves. Svensson and Enli and Skogerbo find that politicians themselves point to deliberation as important motives or reasons for using social media, but that in reality, they do not use deliberation or interactivity as much as they might wish or say they want to use (Svensson 2013, 216, Enli and Skogerbo 2013, 770). They rather find that promotion of themselves and management of their images are motives or reasons more frequently used by the politicians.

1.6. Dependent variable: professionalization in social media profiles

The aim of this master thesis is to measure and explain the level of professionalization in the Twitter profiles of Dutch, Norwegian and Austrian MPs. Adaptation to electronic media such as online websites and social media is one of the strategies used in professionalized political campaigns. But what does a professionalized use of social media entail? As mentioned earlier, professionalization can mean being professional, or using something in a professional way as in not being amateurish, but knowing how to handle the thing in question according to a set (high) standard. In this case, it would be knowing how to handle a social media profile professionally; that there is consistency in the design of the profile and that the design gives the impression that it has been made by someone who knows what he or she is doing, someone who is professional in social media.

As mentioned, an increased focus on the individual politician, or a process of personalization, is also a part of professionalization. Being professional in social media use can also mean, in addition to the more non-amateurish use (or advanced use as I will call it from now on), being personal. The level of professionalization in the MPs social media profiles is the dependent variable of this study, and it is measured as the level of a professional, non-amateurish, or advanced use, and as

personalization, when politicians show the private side of themselves on social media. In this way, there are two “pillars” of professionalization I want to investigate and explain, which forms two types of professionalized Twitter user: the non-amateurish, advanced user, and the personal user.There are other features of professionalized political communication that are possible to focus on and

investigate, however, this is beyond the scope of this study.

1.7. Independent variables: individual and contextual variables: Hypotheses 1.7.1. Individual variables:

In order to explain the levels of professionalization in the Twitter profiles of Dutch, Norwegian and Austrian MPs, I need to find possible independent variables. I follow Larsson and Kalsnes (2014) and distinguish between individual and contextual variables. I further divide these into micro/individual-level, meso/party-micro/individual-level, and macro-level /country variables. Micro or individual variables that could influence the degree of professionalization of the MPs’ Twitter profiles could be individual

characteristics such as age, gender, and ethnicity. Influential contextual, or meso, variables could be characteristics such as party ideology, party structure, party size. A macro-level variable could be country, as MPs in different countries will be influenced by the country they live in and the political climate there.

When it comes to the first individual variable, age, it can be expected that younger MPs are more likely to adopt and use social media (as there are more younger than older users of social media, see source in the introduction), and therefore to be more familiar with professional strategies used here, and devote more time to their social media avatar. Larsson and Kalsnes (2014), when it comes to social media adoption, do indeed find a higher rate of adoption and use of the social media among younger politicians in their study of Norwegian and Swedish politicians. Other researchers have also found similar results for websites (Carlson and Djupsund 2001, Carlson and Strandberg 2005) Therefore, the following hypothesis is formulated regarding age:

H1: The level of professionalization is expected to be higher in the Twitter profiles of younger MPs than in the profiles of older MPs.

Other possible individual explanatory variables could be gender or ethnicity. There have been studies that have investigated differences in campaign strategies among men and women, and found

significant differences (Carlson 2007, Kahn 1996, Puopolo 2001, as paraphrased in Larsson and Kalsnes 2014, 4). However, there have also been studies that have not found any significant

differences. A very relevant study for my thesis that does focus on social media is that of Vergeer et. al (2012). While the Kahn´s and Puopolo´s studies focus on the United States, a country with a very different political system and media system from my case countries of the Netherlands, Norway and Austria, Vergeer et. al. study the Netherlands. Vergeer et. al. (ibid.) find no great differences between male and female Twitter adoption. Also, Kofoed (2015) in her study on the level of professionalization in Dutch MPs social media profiles finds no great differences between male and female MPs. Due to the contradicting evidence found on this point, the following hypotheses were formulated:

H2a: The level of professionalization of Twitter profiles is expected to be higher in the Twitter profiles of male MPs than in the profiles of female MPs.

H2b: The level of professionalization of Twitter profiles is expected to be higher in the Twitter profiles of female MPs than in the profiles of male MPs.

H2c: The level of professionalization of Twitter profiles is not expected to differ significantly in the Twitter profiles of male and female MPs.

There have also been studies on to what extent ethnic minorities have adopted and are active on social media. Spierings and Celis (2014) investigate whether ethnic minorities and women differ from ethnic majority politicians and men in adopting social media profiles and the number of tweets they send out, and their results show that ethnic minority men and women of the ethnic majority were more present on social media, and that men post most tweets. On the other hand, they find that the median number of tweets is higher for ethnic minorities than for ethnic majority candidates (ibid.). Yet, it must be noted that all these differences are quite small. In a study of the level of

professionalization of Dutch MPs, Kofoed (2015) finds no great differences between ethnic majority or ethnic minority MPs. I thus do not expect there to be any great differences here, but include this variable as the level of professionalization may be different in Norway and Austria than in the Netherlands among ethnic majority and minority MPs. Thus, the following hypotheses were formulated about ethnicity:

H3a: The level of professionalization of Twitter profiles is expected to be higher in the Twitter profiles of ethnic majority MPs than in the profiles of ethnic minority MPs.

H3b: The level of professionalization of Twitter profiles is expected to be higher in the Twitter profiles of ethnic minority MPs than in the profiles of ethnic majority MPs.

H3c: The level of professionalization of Twitter profiles is not expected to differ significantly in the Twitter profiles of ethnic majority MPs and ethnic minority MPs.

1.7.2. Contextual variables: meso- and macro-variables

A meso, and contextual, independent variable is that of party size. This could prove an important contextual variable that explains the level of professionalization in the profiles of Dutch, Norwegian and Austrian MPs. This variable is connected to the debate of normalization versus equalization. These theories focus on the role of social media in changing the power dynamics of political systems (Spierings and Jacobs 2014, 2). The normalization thesis assumes that new media will tend to “mainly benefit the parties that were over already dominant”, while the equalization thesis assumes social media will “level the playing field” (ibid.). In the first case, “new technologies will merely reinforce existing inequalities” (Jackson and Lilleker, 2011, as quoted in Spierings and Jacobs, 2014) as bigger parties have more money to hire expert crew to handle social media accounts (Spierings and Jacobs

2014, 15). In other words, big parties have more resources to ensure a professionalized political communication, thus having more advanced user profiles or being more personal in their social media use. In this way, bigger parties will attract more voters, and their dominant position will tend to be reinforced. In the case of equalization, you could argue that smaller parties can gain easy access to social media as adopting a social media account is very cheap, in addition to it being easier to get more media exposure by contacting a journalist on Twitter (ibid., 14, 17).

Research that has been done on this field has shown different results. Spierings and Jacobs (ibid.) conducted research on the social media adoption and presence of Dutch politicians, and conclude that we might see equalization and normalization processes happening at the same time, or perhaps a process where we can first observe equalization, then normalization (ibid., 19). Vergeer and Hermans (2013, 414, as paraphrased in Spierings and Jacobs, 2014, 3) suggest a process of

normalization as the notion that smaller parties will use social media more is not supported in their research. Kofoed (2015) in her study on Dutch MPs professionalization levels on Twitter found that bigger parties had higher levels of professionalization than smaller (i.e. medium-sized) parties, thereby presumably giving them an advantage to attract more voters. Due to the contradicting results of previous research, the following hypotheses were formulated:

H4a: The level of professionalization of Twitter profiles is expected to be higher in profiles belonging to MPs from bigger parties than in profiles belonging to MPs from smaller parties.

H4b: The level of professionalization of Twitter profiles is expected to be higher in profiles belonging to MPs from smaller parties than in profiles belonging to MPs from bigger parties.

The ideology of the MP party might also be of importance. Vergeer et. al. (2011, 485) assume that politicians of progressive parties, i.e. parties that want a change in “the role of the state and the size of the welfare state” unlike conservative parties, will tend to adopt social media profiles faster than politicians from other parties. This is due to due progressive parties’ open-mindedness to change, making them more likely to adopt to innovations like social media (ibid., 491). Their (ibid.) findings show that progressive parties do use social media more (more tweets) than conservative and center parties.

I would suggest to rather use the term “post-materialist” parties than progressive parties as this is a more fitting concept. Post-materialist parties are parties representing postmodern, or

post-materialist, values rather than materialistic, authoritarian, or conservative values. According to Inglehart (2000, 223), postmodern values are present in the younger generation, and imply a greater

focus on “self-expression instead of deference to authority” unlike materialist or conservative values that give priority to economic and physical safety and authority. Postmodernalists focus more on the environment and on sexual freedom, see diversity in the population as interesting, not threatening, and have less interest in absolute and rigid religious beliefs. Examples of parties that represent such values could be green parties or progressive liberal parties. Yet, findings from Kofoed’s (2015) study show that MPs from more established Dutch parties, and particularly from the conservative liberal party of VVD, used a strategy of professionalization more often than MPs from postmodern parties.

Due to these contradicting claims, three hypotheses are formulated:

H5a: The level of professionalization of Twitter profiles is expected to be higher among MPs belonging to conservative parties than among MPs belonging to post-materialist parties. H5b: The level of professionalization of Twitter profiles is expected to be higher among MPs belonging to post-materialist parties than among MPs belonging to conservative parties.

H5c: The level of professionalization of Twitter profiles is not expected to differ to a great extent between MPs belonging to post-materialist parties and MPs belonging to conservative parties.

A further possible ideological variable could be whether a party is populist, as suggested by Jacobs and Spierings (2015). Gibson and Rommele (2009, 280) point to internal centralization as an important variable in explaining professionalization in electoral campaigning, where parties with a more top-down structure and hierarchical culture are expected to engage in professionalized

campaigning as they are more likely to adopt business-like strategies of professionalized campaigning. In this way, you would expect populist parties to display greater levels of professionalization, as they do have a top-down structure and hierarchical culture. However, findings from Kofoed’s (2015) study showed that populist parties in the Netherlands show a low level of professionalization in their social media profiles. How do we explain this? Dolezal (2015, 108) argues that populist parties like

communists and left-socialists will show low levels of professionalization in their online political communication due to their focus on internal cohesion. In this way, the explanation could be that in populist parties, the leader is the party. He is the leader internally and externally, and is the face of the party to the public. In this way, it can be expected that other MPs from the populist parties will not have very professional social media profiles, as they should not “steal the shine” of the leader. Thus, the following hypothesis was formulated:

H6: The level of professionalization of Twitter profiles is expected to be lower among MPs from populist parties than among MPs from mainstream parties.

When looking at macro-variables, the obvious one, except for the already mentioned media system and political system, is country. MPs might have different levels of professionalization in their Twitter profiles according to which country they live in. The political environment in one country might put a greater focus on professionalized political communication than in other countries. Due to the lack of comparative studies that examine the level of professionalization among politicians from different countries, there is little evidence to back a hypothesis of MPs from which country should have more or less professionalized Twitter profiles. However, I see it as likely that the levels of professionalization is about the same among Dutch and Norwegian MPs as there is about the same amount of people active on social media in the Netherlands and Norway. As there are quite many active on social media in these countries, politicians here will have a great incentive to adopt social media profiles and to make an effort into their profiles, thus making them more professionalized. However, I expect Austrian MPs to have a lower level of professionalization in their Twitter profiles compared to Norwegian (and Dutch) MPs as there is a lower number of the population active on social media in Austria than in the two other countries. If there are not so many people using social media, politicians will not have a great incentive to register at an SNS and to make an effort into having a good, or professionalized, profile. From this, the following hypotheses are formulated about the country variable:

H7: The level of professionalization in the Twitter profiles of Dutch and Norwegian is not expected to differ to a great extent.

H8: The level of professionalization is expected to be lower among MPs from Austria than among MPs from Norway and the Netherlands.

Chapter 2: Methodogical framework

2.1. Case selection

2.1.1. Most similar systems design and consensus democracies

The countries I have chosen as cases for my study are the Netherlands, Norway and Austria. One reason why I have chosen these cases is due to the fact that they are continental European countries. The study of the professionalization of political communication, of the personalization of politics, and of social media and politicians have long been dominated by researchers from Anglo-Saxon countries, focusing primarily on Anglo-Saxon cases (Bennett 2012, McAllister 2005, Farrell et. al. 2001,

Harfoush 2009, Gulati and Williams 2007, Norris 2003). Anglo-Saxon countries, especially the USA, are frequently candidate-centered political systems, many of them presidential. The countries in my study are more party-centered, parliamentary systems where you would not expect as much focus on the individual politician and not such a high degree of professionalization of political communication as in Anglo-Saxon countries. I therefore think it could be valuable to study these countries in order to establish the degree of professionalization of political communication in party-centered, parliamentary European countries.

In addition to this, my study is of a comparative nature. Many studies on professionalization in social media have been case studies, only studying one country at the time (Svensson 2013, Enli and Skogerbo, Spierings and Celis 2014, Vergeer et. Al. 2013), thereby missing the opportunities comparative studies have for greater nuance and insight. Many social media studies also mainly focus on whether politicians are present on social media or not, or are concerned with tweets, whether it be the number of tweets or the content of these (Vergeer et. al. 2013, Hegelich and Shahrezaye 2014, Larsson and Kalsnes 201). Few studies have focused on how politicians present themselves in their profile information and their profile pictures; whether these give the impression of a professional, advanced user profile, whether it gives a more personal impression of the politician, etc. The aim of my study is the latter, thereby giving an original contribution to the study of professionalization of political communication and political social media use.

My three cases are similar in several ways. Thus, I have chosen these cases according to a most-similar systems design as they are similar on the measured independent variables (Seawright and Gerring 2008). Yet, they might differ on one independent variable, which can provide a basis for explanation of the Y, the outcome. All the chosen cases are, as mentioned, party-centered,

parliamentary political systems, they are examples of what Lijphart would call consensus democracies, unlike Anglo-Saxon countries that are often characterized as majoritarian democracies. In consensus democracies government power is shared between two or more parties, there is an

executive-legislative balance of power, multiparty systems, proportional representation, corporatist interest groups aimed at compromise, decentralized government, division of legislative power between two parliamentary houses, rigid constitutions, a system where supreme courts review laws on whether they are constitutional, and independent central banks (Lijphart 1999, 3-4). This is in contrast to

majoritarian systems that have a concentration of power in single-party governments, a dominant position of the executive in relation to the legislative, a two-party system, a majoritarian electoral system, pluralist interest groups that work in a system that follows a rule of free-for-all competition among many groups, have a centralized government, a concentration of legislative power in a

parliament of only one chamber, flexible constitutions that can easily be changed, a system where the legislature has the final say in whether their own laws are constitutional or not, and a system where central banks are dependent upon the executive (ibid.). Adam and Meier (2010, as paraphrased in Hermans and Vergeer 2012, 6) argue that presidential systems provide better conditions for personalization than parliamentary systems as the former are more focused on persons rather than parties as the latter, although personalization can also be present in the parliamentary systems. Likewise, majoritarian systems are thought to provide better conditions for personalization than consensus democracies due to the great amount of power the leader has in these countries (ibid.). The advantage of choosing a most-similar-systems design is that as my cases are all typical cases in that they are representative for the rest of the population, we can expect that the results on

professionalization in social media profiles that I find in this study can be generalized to the wider population (Seawright and Gerring 2008, 298), i.e. to other consensus democracies. Also, in this research design you can have many cases, thus escaping the problems small-N cases might encounter. However, it can be difficult to measure scalar independent variables like wealth of a country in most similar cases as no countries have exactly the same score on these dimensions (ibid., 305).

2.1.2. The role of electoral systems for personalization

When it comes to factors that might explain the level of professionalization in the social media profiles of Dutch, Norwegian and Austrian MPs, the electoral system in a country plays an important role in this regard. I took Josep Colomer’s (2011) framework on the impact of electoral systems on different degrees of party or personal representation as a starting point when choosing the cases of the

Netherlands, Norway and Austria for this study. All the case countries have the same type of electoral system and a similar ballot structure. They can therefore be said to be similar on these factors, making them most similar system cases with findings generalizable to other similar cases.

Party representation is political representation of citizens by parties, while personal representation is the representation of citizens by individual politicians (ibid.). Colomer (ibid., 7) refers to personal representation as “the personal quality of representatives, that is, their reliability and ability to fulfill electoral promises and respond to voters’ demands”. In party representation on the other hand, the individual politician and even the party leader is less important than in personal representation, as the “glue” between the politician and the citizen is (party) ideology, not primarily the qualities of the individual politician. Colomer (ibid.) comes up with a new classification of voting procedures that influences whether we will tend to see more personal or more partisan representation, based on two elements: firstly, the number of choices available to the voter, and secondly, the

opportunity to vote for either political parties or individual candidates. Following these elements, we get three types of ballot structure: closed, semi-open and open (ibid.).

In the closed ballot structure, the voter only has one choice to make, namely to vote for a party. The voter cannot decide between different candidates from the same party when voting. This is used in single-seat systems, for example in Britain, in some mixed systems like Mexico, and in closed party list systems like in Spain. This ballot form produces party representation, it does not give the incentives for individual candidates to reach out to voters on the ground of their persona (ibid., 15). In the semi-open ballot form, the voter has more than one choice in elections, as he or she can vote for both a party and a candidate. Examples of such a system could be the two-round system in France where the voter is given two choices, but at two different times, one choice for a candidate and one for a party. The double vote system, as used for example in Germany, gives the voter a choice for a party and for a candidate who may not be representing the party he or she voted for. In preferential list systems (as used in the countries of my study), the voter can choose between a party and one or more candidates from the same party. In the open ballot system, the voter has more than two choices when choosing party and individual candidates. Ireland, for example, uses ordinal rank ballot where voters can list the candidates according to which ones they like best and least, and the candidates on the list can be from different parties. In Switzerland, the ballot system allows the voter to choose as many candidates as there are seats to be filled in the election, regardless of the party the candidates are members of.

Colomer (ibid., 9-10) notes that the three categories are compatible with “majority”, “mixed”, and “proportional” rules, which gives us a two-dimensional issue with nine categories. The

assumption is that politicians in closed ballot systems will not have a great incentive to have close relations to their voters, as voters only vote for parties, and cannot choose between different candidates (ibid.,15). In semi-open and especially open ballot systems with proportional rules,

however, politicians will have a greater incentive to have close relation with their voters, as the voters can choose between candidates, not only between parties (ibid.).

The cases I chose for this study, the Netherlands, Norway and Austria, all have semi-open ballot systems with preferential votes. This should mean that MPs in these countries will have a medium level of incentive to have close relationships with their voters, which will show itself as a medium high level of personalization in their Twitter profiles. Colomer (ibid., 16) argues that there “is a tendency towards increasing numbers of preferential votes in successive elections” in countries like Austria and the Netherlands. He does not mention Norway, but he does mention Denmark and Sweden which Norway is often regarded to be very similar to when it comes to the political and ballot system, arguing that the amounts of preferential votes are stable (ibid.). To decide whether or not Norway (or any of the other case countries) could be thought to be less or more personalized than the other cases according to its electoral system is, however, not the purpose of this study. I assume that if there are differences between the countries when it comes to level of personalization due to electoral systems, these will be rather small. I will thus work from the assumption that these countries are most similar cases, they have more or less the same level of personalization due to the similarities of their electoral systems (and other characteristics), and that we thus can see roughly the same level of personalization in the profiles of Dutch, Norwegian and Austrian MPs.

2.1.3. The role of the media system for professionalization

A variable that might explain the level of professionalization in the social media profiles of politicians is the media system of a country. I argue, following Strandberg (2008), that what kind of

characteristics a media system in a country possesses profoundly influences how politicians

communicate with their electorate in the media, including the strategies they use for communication in social media. I used Hallin and Mancini’s typology of media systems as a starting point for

distinguishing between different media systems, and for choosing the cases in this study. All the three cases of the study are classified as having the same media system, which contributes to making them most similar system cases in addition to the other characteristics they have in common.

Hallin and Mancini (2004) outline three media models according to both political and media characteristics in different countries, and also correspond to geographical areas. The first is a

Mediterranean or Polarized Pluralist Model which is characterized by a press oriented towards the elite in the country, a small paper circulation and centrality of electronic media (ibid., 73). Press freedom developed late here. There is a low level of professionalization of journalism and a high level of political parallelism as the press has been strongly marked by instrumentalization by the government and political parties. The state plays an important role as it is often the owner, regulator and funder of the media. The Polarized pluralist model is prevalent in countries like France, Greece, Italy, Spain and Portugal.

The second model is an Anglo-Saxon, or Liberal, media model (ibid., 75). Press freedom developed early in these countries, as did the mass-circulation press. Commercial newspapers are common, and the role of the market is generally quite strong, with a limited role for the state. Political parallelism is low and internal pluralism high, as is the professionalization of journalism. Countries in which this model is common is Britain, the USA, Canada and Ireland.

The third model, the North/Central European or Democratic Corporatist Model is predominant in the three case countries I have chosen: the Netherlands, Norway and Austria (as well as in other central continental and Nordic countries) (ibid., 74). This model is characterized by early development of freedom of press, a very high newspaper circulation, high journalistic professionalization, a strong role for the state and strong political parallelism. Many newspapers used to be affiliated with or controlled by political parties or organized social groups, though this characteristic has become less important over time. Public broadcasting has been influenced by the segmented nature of society, each social group controlling its broadcasting station. We have also seen stronger commercialization of the press the last decades.

Hallin and Mancini (ibid., 76) argue that the media in the Liberal countries are closer to business than to the political world, while it is the other way around in the Polarized Pluralist

countries, while the Democratic Corporatist countries are somewhere in between. In this way, we can expect that in the Liberal media model, politics and political communication will be more

commercialized as the media is highly commercialized. In this way, we could also expect a high level of professionalization in these countries, as the processes of mediatization and professionalization of political campaigning can be assumed to have progressed the farthest here. Countries such as the Netherlands, Norway and Austria are characterized by a medium high level of media

commercialization, and we might thus see a medium high level of professionalization in political communication, also in politicians’ social media use.

It is thus expected that the Netherlands, Norway and Austria are, according to both their ballot structure and media system, expected to show medium high levels of professionalization in political communication, which is the reason why I have chosen these three countries as cases. It can also be expected that we can generalize the findings of this study from these three countries to other countries that have the similar ballot structure and similar media characteristics.

2.1.4. Characteristics of the three case countries – parties in parliament and social media penetration

To summarize the relevant characteristics of the three case countries, I have constructed a table that provides an overview of these.

Table 1 – Case characteristics

Cases Party system Political system No. of elected parties in parliament Current government No. of MPs in parliament The Netherlands Multi-party system Constitutional monarchy 11 VVD-PvdA 150 Norway Multi-party system Constitutional monarchy 8 Høyre-FRP 169 Austria Multi-party system Parliamentary republic 6 SPÖ – ÖVP 183

Cases Ballot structure Media system Monthly social media penetration in population, in %* Twitter penetration among MPs, in %

The Netherlands Semi-open Democratic corporatist

52 97

Norway Semi-open Democratic

corporatist

64 77

Austria Semi-open Democratic

corporatist

39 38

*Statistics collected from Statista, the Statistics portal 2015.

As previously mentioned, all countries are parliamentary systems (the Netherlands and Norway are constitutional monarchies, Austria a parliamentary republic) with a multi-party system and

proportional representation. The Netherlands have the highest number of parties in government (due to no existing electoral threshold), currently with 11 parties, followed by Norway with 8 parties, and